On May 16, Sam Altman testified before the US Senate Judiciary Committee about artificial intelligence. It was Altman’s debut performance on Capitol Hill, the first time he’s testified about his work as the CEO of OpenAI, the company responsible for the large language model GPT-4 and its most famous consumer application, ChatGPT.
Congressional tech hearings are typically antagonistic, especially those involving social media, ever a thorn in the side of federal lawmakers. “Congress failed to meet the moment on social media,” said senator Richard Blumenthal, the subcommittee chair hosting Altman. “Now we have the obligation to do it on AI before the threats and the risks become real.”
While lawmakers expressed deep concern about the development of AI, they largely treated Altman like a responsible steward of an inevitable technology rather than an adversarial industry leader. Washington Post reporter Cristiano Lima called the hearing “much less combative” than the Congressional debuts of Meta CEO Mark Zuckerberg, Alphabet CEO Sundar Pichai, Amazon founder Jeff Bezos, or Twitter co-founder and former CEO Jack Dorsey.
“The treatment Altman’s getting is night and day from what social media execs have gotten,” tech journalist Issie Lapowsky wrote on Twitter, noting it was one of Atlman’s fellow panelists, the New York University professor Gary Marcus, who pressed Altman to answer one of Blumenthal’s questions about a worst-case scenario for AI.
“My worst fear is we cause significant harm to the world,” Altman said. “If this technology goes wrong, it can go quite wrong.”
Sam Altman to Congress: Please regulate us
During the course of the hearing, Altman did something that many other tech executives have done, calling for regulation of his industry. It’s a smart maneuver: If you’re going to be regulated, you might as well seem welcoming to change and help mold it.
For his part, Altman proposed a regulatory regime involving a brand new agency to regulate AI and license the development of large-scale models, as well as independent auditing and testing requirements. (It’s not clear how licensing requirements would jibe with the First Amendment, which protects computer code as speech.)
There wasn’t a clear path forward to regulation by the end of the hearing, though Blumenthal noted it would be the first in a series on AI. But it was clear that Altman had won the day, emerging from a trip to Washington largely unscathed—if not empowered.
On May 15, the day before the hearing, Altman reportedly spoke to a group of 60 lawmakers at a private dinner. CNBC reported that Altman wowed the crowd in a two-hour spiel, winning rave reviews from seasoned legislators. It’s not clear if the night-before dinner helped Altman’s rapport with the Senate, but it surely didn’t seem to hurt.